Can Cities Add More Computing Power Without Adding More Environmental Pressure? China Offers New Evidence

Can Cities Add More Computing Power Without Adding More Environmental Pressure? China Offers New Evidence
Representative image. Credit: ChatGPT

The digital economy is expanding on infrastructure that is largely invisible but increasingly crucial to both economic growth and environmental pressure. Cloud computing, artificial intelligence, digital payments and online public services all require more servers, storage and electricity, forcing governments to confront a harder question: how can computing capacity expand without locking cities into higher energy use and emissions?

A study titled "Green Data Center Pilots and Urban Coordinated Development of Digital and Green Transformation: Evidence from China," published in Sustainability by Yang Zhang and Yongfeng Hou, examines that challenge using data from 284 Chinese cities between 2011 and 2023.

The researchers find that participation in China's National Green Data Center programme was associated with about a 5.86% improvement in coordinated digital and green development, suggesting that the way computing infrastructure is designed and connected to industry and policy could shape whether digital expansion supports or strains environmental goals.

Digital Growth Is Creating an Infrastructure Constraint, Not Just an Energy Bill

Data centers have become foundational infrastructure for modern economies, yet their environmental profile remains difficult to ignore. They operate continuously, consume large volumes of electricity and can increase emissions where power generation remains dependent on fossil fuels. Equipment production, replacement and disposal add another layer of environmental cost.

Green data centers attempt to change the operating model through more efficient servers and cooling systems, intelligent workload management, renewable electricity and better coordination of computing resources. China had 246 national-level green data centers in operation by the end of 2024, alongside more than 300 types of energy-saving and carbon-reduction technologies deployed across data centers, telecom base stations and related infrastructure, according to figures cited in the study.

The researchers move beyond measuring whether individual facilities save electricity. Their index tracks whether digitalization and environmental performance improve together across entire cities. Digital indicators include telecommunications activity, broadband and mobile access, digital-sector employment and policy support, while green indicators cover pollution control, public services and urban ecological conditions.

After controlling for economic development, population density, wages, openness, public expenditure, education and environmental spending, exposure to the green data center programme was associated with a 0.0186 increase in the digital-green coordination index, equivalent to about 5.86% of the sample mean. Separate estimates also showed statistically significant improvement in both digitalization and greenization.

Three results stand out:

  • Green data center exposure was associated with stronger digital and green innovation activity.
  • Pilot cities recorded greater concentration of 5G, industrial Internet and blockchain-related firms.
  • Urban energy-use efficiency also improved, suggesting that computing expansion can become less resource intensive when infrastructure and management systems improve together.

China's Experience Suggests Data Centers Can Become Industrial Infrastructure

A data center can be treated as a large electricity-consuming building, or it can be treated as a shared economic platform. The distinction is important. The study argues that centralized computing capacity can reduce the need for firms to build fragmented in-house systems. Access to shared storage, processing and cloud services can lower infrastructure costs for data-intensive businesses, while more efficient equipment and workload management can reduce the energy required to provide those services.

The empirical evidence is consistent with this wider industrial role. Green data center participation was positively associated with both green and digital patent activity, while pilot cities also experienced stronger relative agglomeration of computing-related industries. The authors do not describe these tests as conclusive proof of causal mediation, but they offer plausible pathways connecting infrastructure investment to wider urban transformation.

For emerging economies, this has practical consequences. A data center strategy focused only on construction spending, land allocation or nominal capacity can produce expensive infrastructure without generating substantial local economic spillovers. Stronger returns are more likely when computing facilities connect directly with firms, public services, innovation ecosystems and growing digital applications.

The pattern across sectors reinforces that point. The study found stronger effects where green data centers were linked to telecommunications and Internet activities, sectors that use computing capacity intensively and continuously. The incremental effect was weaker for public institutions, while the additional effect for financial-sector pilots was not statistically significant.

Infrastructure, in other words, appears to create larger development gains when there is enough real demand to absorb it.

Policy Coordination Matters More Than Simply Adding More Green Programmes

What happens when green data centers overlap with other policies? The researchers examined interactions with Broadband China, artificial intelligence pilot zones, green finance pilots and zero-waste city programmes. Green data centers produced statistically significant positive interaction effects with the broadband and AI initiatives. No statistically identifiable additional interaction effect appeared with the two green-oriented programmes during the study period.

The authors interpret the contrast through functional complementarity. Broadband policies expand connectivity. AI programmes increase demand for computing and data-processing services. Green data centers provide the energy-efficient computing capacity needed to support those activities. Each policy therefore addresses a different part of the same system.

Policies with similar environmental objectives may instead overlap, producing less additional benefit than governments expect simply from implementing several programmes in the same place.

The lesson is particularly relevant to countries where digital, industrial, energy and climate strategies are developed by different ministries. Green computing cannot be optimized only through environmental standards, while digital policy cannot treat electricity supply as an external concern.

Governments may need to plan data-center expansion alongside broadband deployment, clean-power availability, industrial demand and regional infrastructure. The study also found stronger effects in cities covered by national urban-agglomeration policies, suggesting that cross-city connectivity and regional coordination can help spread computing resources, technologies and energy benefits beyond a single jurisdiction.

The Next Race Will Be About Efficiency, Location and Demand

The research offers a compelling direction, but not a universal formula. Its evidence comes from Chinese cities operating within a specific institutional, energy and policy environment. The authors acknowledge that unobserved time-varying factors cannot be eliminated completely and that results may not transfer directly to countries with different regulatory systems, power markets or levels of digital development.

Another unresolved issue is geographical displacement. A city may improve the efficiency of its data centers while drawing electricity generated elsewhere, potentially shifting emissions or environmental pressure rather than eliminating them. The study does not explicitly trace these cross-regional effects, an omission the authors identify as a priority for future research.

Those limitations are especially important as AI accelerates demand for computing infrastructure. Countries will increasingly compete to host data centers, cloud regions and AI facilities, but headline capacity figures reveal little about whether those investments strengthen domestic industries, increase energy productivity or create new environmental vulnerabilities.

Overall, China's green data center experiment suggests that computing infrastructure can do more than process information. Designed well, it can become part of an urban development system linking innovation, energy efficiency, industrial concentration and environmental performance. Designed poorly, the same expansion risks producing more capacity without comparable gains in productivity or sustainability.

  • FIRST PUBLISHED IN:
  • Devdiscourse
Give Feedback

Use this form for editorial or site feedback. We usually reply within 2 to 3 working days.

By submitting, you agree that we may use your email address to respond.